Can Trust in Artificial Intelligence Be Measured?

On 8 September 2026 the Know Center from Graz publicly launched a new quality seal called “Trustworthy AI – by Know Center” at a press conference in Vienna. The institute presented the seal as the start of a structured assessment scheme intended to evaluate AI components and systems against a set of technical and qualitative criteria. According to the Know Center’s statements and accompanying press materials, the initiative aims to make aspects of algorithmic trustworthiness both measurable and communicable.

At the same event the institute announced the first successful completion of its qualification process: the prototype “Kommunaler KI‑Assistent”, developed by IT‑Kommunal for the Austrian Association of Cities (Österreichischer Städtebund). The announcement was echoed in the communications of the cities association and the implementing company, which framed the prototype as a lighthouse project to demonstrate the application of the new assessment methodology.

According to the Know Center, the seal relies on a two‑phase assessment: an initial technical audit followed by a qualitative review. The assessment framework covers transparency, fairness, robustness, security, human oversight, privacy, and digital sovereignty. Upon successful evaluation, a certificate with a unique Seal‑ID is issued; the validity of a seal is limited to 24 months, the institute states.

The description on the Know Center website indicates the process includes both technical tests—such as checks for robustness and traceability of decisions—and organizational documentation, including governance processes, data‑protection measures, and steps taken to prevent discrimination. Thus the seal is positioned deliberately beyond simple performance benchmarks: it bundles multiple criteria to reflect the broader notion of ‘trustworthiness’ rather than single metrics like accuracy.

The first certified system is the so‑called “Kommunaler KI‑Assistent”, a prototype intended to support municipal staff by providing information, suggestions and administrative assistance. Developed by IT‑Kommunal on behalf of the Austrian Association of Cities, the prototype is presented as an example of how the seal can be applied to public‑sector IT services that interact directly with citizens.

Choosing a municipal assistant as the inaugural recipient sends a dual signal: it highlights the relevance of trustworthy AI for public administration—where accountability and legal clarity are particularly important—and raises practical questions about deployment. Integrating such a system into existing municipal IT ecosystems, funding procurement and maintenance, and allocating legal responsibilities in case of erroneous automated recommendations will be practical issues for municipalities to address.

The Know Center frames its seal as a scientifically based assessment tool; it does not claim legally binding regulatory authority. At European level, the AI Act establishes mandatory compliance requirements for certain high‑risk systems and sets out formal conformity assessment procedures. How the Know Center’s voluntary seal will interact with or complement statutory conformity checks under the AI Act is not yet determined: it currently functions as a quality marker rather than a substitute for legally required certifications.

A key open question is whether public authorities, regulators or procurement bodies will accept the seal as evidence of compliance or quality in official processes. The seal could be used instrumentally as supporting documentation in tenders or internal audits, but unless national authorities or EU bodies formally reference it, its legal standing will remain limited. The path towards institutional recognition will depend on transparency of the assessment process, oversight mechanisms, and possibly on third‑party audits.

The new seal offers potential benefits: it gives developers and operators a structured route to demonstrate responsible practices, and it may raise user awareness about the dimensions that constitute trustworthy design. For municipalities and other public actors, a trusted assessment framework could help standardize procurement and provide citizens with clearer information about the safeguards in place.

However, multiple practical challenges persist. The scalability of a two‑phase audit process under rising demand is uncertain; certification costs might be a barrier for smaller vendors or municipalities. Governance questions also arise: how to ensure the independence of assessments, how to handle conflicts of interest if research organizations both develop and certify systems, and how to prevent the seal from becoming a marketing label with limited substance. The Know Center has published criteria and procedures, but issues of recognition, funding, oversight and long‑term sustainability remain unresolved. The answers to these operational and governance questions will determine whether the seal becomes a useful tool in practice or stays primarily a research‑driven standard of best practice.